+++ /dev/null
-2011-08-22 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (create_rrd_graph): only use daemon if socket
- exists.
-
- * PVE/pvecm (delnode): test if we have quorum.
-
- * src/cfs-plug-link.c (cfs_plug_link_getattr): set read-only if we
- do not have quorum.
-
-2011-08-19 Proxmox Support Team <support@proxmox.com>
-
- * PVE/pvecm (create): removed section (rm) from cluster.conf (do
- not start rgmanager by default).
-
- * PVE/Cluster.pm (check_cfs_quorum): use 'local' instead of
- 'prv/lock', because this link always exists on the file system.
- (ipcc_send_rec_json, ipcc_send_rec): fix error check
-
-2011-08-18 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (ssh_merge_keys): add check if
- /etc/pve/priv/authorized_keys exists
-
-2011-08-16 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (parse_datacenter_config)
- (write_datacenter_config): register reader/writer for
- datacenter.cfg
-
-2011-08-15 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm: add 'datacenter.cfg' to observed files.
-
-2011-08-10 Proxmox Support Team <support@proxmox.com>
-
- * PVE/pvecert: removed binray (use pvecm instead)
-
- * PVE/pvecm (lsnode): remove verifyapi handling code.
- (updatecerts): new method (replacement for pvecert binary)
-
- * PVE/pvecert: also update ssh related files.
-
- * PVE/Cluster.pm (ssh_merge_known_hosts): replace entries with
- wrong key. new parameter createLink.
-
-2011-08-09 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (ssh_merge_known_hosts): also manage known_hosts
- file.
-
- * PVE/pvecm (backup_database): revert pervious change - do not
- pass IP to addnode. Instead we run gen_pve_node_files in 'add'
- after we get quorum.
-
-2011-08-08 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (check_cfs_quorum): check for undefined $st (no
- such file).
-
- * PVE/pvecm: try to create all files/dirs in 'addnode', because that
- node has quorum.
- (create): Add id_rsa.pub to authorized_keys.
- (create): Wait for quorum and restart apache2
-
- * src/cfs-plug-func.c (cfs_plug_func_new): add write callback (to
- trigger actions).
-
- * src/pmxcfs.c (write_debug_setting_cb): implement callback to
- enable/disable debug mode (.debug file)
-
-2011-08-05 Proxmox Support Team <support@proxmox.com>
-
- * PVE/pvecm (create): call keygen with parameter hash
-
-2011-08-04 Proxmox Support Team <support@proxmox.com>
-
- * src/pmxcfs.c (create_symlinks): remove status dir and link
-
- * src/*: remove syslog_cache (not needed)
-
-2011-08-03 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-func.c: simple plugin to view data returned by
- function. We use that to display status inside the filesystem
- (.version, .members, .vmlist, .rrd, .clusterlog)
-
-2011-08-02 Proxmox Support Team <support@proxmox.com>
-
- * src/dfsm.c (dfsm_set_mode): call new callback dfsm_synced_fn()
-
- * src/status.c: add 'iphash' to store ip adresses of all
- nodes.
- (nodeip_hash_set): helper to update iphash
- (dfsm_synced): send ip address when we are in sync
-
- * src/pmxcfs.c (main): set cfs.ip to local ip address.
-
-2011-08-01 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (check_cfs_quorum): helper to check quorum
- (cfs_lock_file): remove $text parameter
- (cfs_lock_storage): remove $text parameter
-
-2011-06-29 Proxmox Support Team <support@proxmox.com>
-
- * PVE/Cluster.pm (run_silent_cmd): supress output when running
- openssl.
-
-2011-05-04 Proxmox Support Team <support@proxmox.com>
-
- * src/status.c (kvstore_send_update_message): use correct data
- pointer (data instead of &data)
-
-2011-03-17 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (get_nodelist): simply return an array of node
- names.
- (get_tasklist): get cluster wide task list
- (broadcast_tasklist): send tasklist to cluster nodes
-
-2011-03-10 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (rrd_dump): cache RRD data for one second (avoid
- to many reads from API server)
- (rrd_dump): parse RRD data
-
- * src/status.c (update_rrd_data): add RRD definition for VMs
- (update_rrd_data): use thread safe API, connect to rrdcached
-
- * src/dfsm.c (dfsm_send_message_full): limit to 100 retries (max
- 10 seconds delay).
-
-2011-03-07 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (ipcc_send_rec*): allow to return 'undef'
-
- * src/status.c (cfs_rrd_dump): impl. rrd dump
-
- * perl/Cluster.pm (broadcast_rrd): start rrd integration
-
-2011-02-25 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (cfs_lock_storage): impl.
-
-2011-02-22 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-utils.c (path_is_private): skip '/'
-
-2011-02-16 Proxmox Support Team <support@proxmox.com>
-
- * perl/pvecm: do not use PVE::RPCEnvironment
- (add): refuse to join cluster if there are existing vms.
-
- * perl/pvecert: do not use PVE::RPCEnvironment
-
-2011-02-15 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (cfs_file_version): A way to get the actual
- version of a file. Can be used to detect file modifications.
-
-2011-02-02 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (remote_node_ip): implement stub
-
-2011-02-01 Proxmox Support Team <support@proxmox.com>
-
- * perl/Cluster.pm (cfs_lock_file): first try.
-
-2011-01-31 Proxmox Support Team <support@proxmox.com>
-
- * src/dcdb.c (dcdb_deliver): implement a simple distributed lock
- manager for short lived locks (no fencing, see README).
-
-2011-01-19 root <root@maui.maurer-it.com>
-
- * perl/Cluster.pm (log_msg): fixes for taint mode (untaint $0)
- (log_msg): allow to use 'symbolic' priority names
-
-2010-11-17 Proxmox Support Team <support@proxmox.com>
-
- * src/cpg.c (dcdb_main_loop): use CPG_DISPATCH_ONE instead of
- CPG_DISPATCH_ALL - which can hang for very long time if there is
- much traffic.
-
- * src/cfs-plug-memdb.c (cfs_plug_memdb_readdir): correctly fill
- stat buffer (avoid extra call to getattr (race condition))
-
-2010-11-16 Proxmox Support Team <support@proxmox.com>
-
- (atomic_write_file): implement our own because
- g_file_set_contents() is clumsy (forks subprocess, can't set mode/owner).
-
-2010-11-15 Proxmox Support Team <support@proxmox.com>
-
- * src/memdb.c: replace g_hash_table_insert with
- g_hash_table_replace (insert does not replace keys!)
-
- * src/memdb-backend-bdb.c (bdb_backend_open): open with DB_THREAD
- (just to be sure - we only access from one thread)
-
-2010-11-05 Proxmox Support Team <support@proxmox.com>
-
- * src/memdb.c (memdb_encode_index): use g_checksum instead of NSS
-
- * src/cfs-plug-memdb.c (heartbeat): touch pid file each 5
- seconds. We can use this to test if the filesysten is up and
- running using inotify.
-
- * src/pmxcfs.c (cfs_fuse_statfs): implemented
-
- * src/cfs-plug.c (cfs_plug_base_*): check for correct plug type
- (plug->ops == &cfs_ops)
-
-2010-11-02 Proxmox Support Team <support@proxmox.com>
-
- * src/memdb.c: implement vmlist magic, add check for duplicate VMIDs
-
-2010-10-15 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-memdb.c: implement cluster/local mode
-
-2010-02-26 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-memdb.c (dcdb_worker_thread): handle
- CS_ERR_TRY_AGAIN in cpg_join
-
-2010-02-23 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-memdb.c (dcdb_worker_thread): cpg_group_name.length
- should include trailing zero.
-
- * src/Makefile.am (CFLAGS): use -std=gnu99
-
- * src/pmxcfs.c (main): wait up to 10 seconds to aquire lock.
-
- * src/cfs-plug-memdb.c (dcdb_sync_cluster_conf): only write
- cluster config when there are changes.
-
-2010-02-19 Proxmox Support Team <support@proxmox.com>
-
- * src/pmxcfs.c (main): use LOCK_NB to avoid blocking
- (main): write pid to /var/run/pve-cluster.pid
-
-2010-02-18 Proxmox Support Team <support@proxmox.com>
-
- * src/pvecm: add 'keygen' command to create corosync keys
-
- * src/pvecm: set authkey to "/var/lib/pve-cluster/corosync.authkey"
-
-2010-02-11 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-memdb.c (dcdb_sync_cluster_conf): only overwrite
- local cluster.conf if version is newer.
-
-2010-02-10 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-memdb.c (dcdb_sync_cluster_conf): sync cluster.conf
- to /etc/cluster/cluster.conf
-
- * src/pmxcfs.c (main): add new '-d' option for debugging (don't
- fork and log to console)
-
-2010-02-08 Proxmox Support Team <support@proxmox.com>
-
- * src/memdb-backend-bdb.c (bdb_backend_checkpoint): checkpoint DB
- every minute.
-
-2010-02-03 Proxmox Support Team <support@proxmox.com>
-
- * src/cfs-plug-memdb.c (quorum_notification_fn): use libquorum
- instead of libvotequorum. This should work with any quorum
- provider (for example cman)
-
- * src/cfs-plug-memdb.c (cpg_deliver_callback): always sync state,
- even if we have no quorum. Only reject writes when we are not
- quorate.
-